ResiliNet: A C++ Socket-Driven Store-and-Forward Mesh Communication Framework for Disaster-Resilient Emergency Response
Year: 2026
Type: preprint
Abstract: Natural disasters such as earthquakes, floods, landslides, and cyclones frequently destroy cellular base stations, sever fiber-optic backhaul links, and overload remaining communication infrastructure. When conventional mobile data and end-to-end Internet connectivity become unavailable, emergency response teams operating in disaster zones lose the ability to transmit rescue requests, GPS coordinates, damage photographs, videos, and tactical documents to central command centers. Conventional client-server messaging platforms fail completely under such partitioned or intermittent network topologies. This paper presents ResiliNet, a disruption-tolerant, disaster-resilient emergency communication platform built on store-and-forward mesh networking principles. Unlike traditional applications that require a continuous end-to-end path to a central server, ResiliNet enables participating field devices to create emergency reports, persist them locally in crash-safe storage, and opportunistically relay them across intermediate nodes whenever transient peer-to-peer links become available. The core communication engine is implemented in C++ using low-level socket programming, providing deterministic control over packet framing, non-blocking I/O, data chunking, checksum integrity verification, selective retransmission, duplicate suppression, and four-tier priority scheduling (Critical, High, Normal, and Low). When a relaying node eventually enters an area with active backhaul connectivity, an edge gateway synchronizes queued messages to a Python FastAPI backend and a React-based Central Command Dashboard, enabling coordinators to monitor live node health, queue backlogs, and verified delivery states. Experimental validation under simulated packet loss (up to 50%), link outages, and multi-hop topologies demonstrates that ResiliNet maintains over 91.8% message delivery ratio where direct TCP clients drop below 10%, while reducing critical emergency alert latency by 86.7% compared to unprioritized FIFO queues.
Institution: MIT Art, Design and Technology University
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7235519773
